Many devices exist today that take voice input via a plurality of microphones oriented about the device. These devices range from conference phones, smartphones and smart speakers to other home appliances. A typical device may include four or more microphones. A device's microphone array may be set to cover all parts of a room. However, as a person moves about in a room, one microphone of the array may better detect that person's voice than the other microphones. Similarly, if there are multiple people using the device simultaneously, different microphones may be better for detecting different people. For the highest quality audio processing, the device should use the microphone that best receives the current input. This may entail strategically switching from one microphone to another.
What is needed is a technique to provide more robust switching among multiple microphones on a device when there are multiple voices and/or a single voice that is moving around.